In today's society, despite of rapid development of the Internet, TV is still the mainstream media platform, and users watch TV programs to understand information disseminated from TV stations. If a user inadvertently watches a segment of a program and wants to continue watching the follow-up program, it is necessary to identify the program.
In a conventional method, the user may acquire a keyword appearing in a segment of a program, manually click on a keyboard or click on a touch screen to input the keyword to a terminal; the terminal acquires the keyword input by the user; sends a program request to a server, where the program request is used for identifying a program corresponding to the keyword; receives a program returned by the server according to the program request, and takes the program as a TV program identification result of the keyword.
In the conventional method, the user cannot acquire a keyword of a program or may acquire a wrong keyword, so that the program request sent by the terminal to the server according to the keyword is wrong, and a correct program cannot be identified, which reduces accuracy of TV program identification.